ARTICLE Received 25 Jun 2015 | Accepted 23 Dec 2015 | Published 28 Jan 2016
DOI: 10.1038/ncomms10541
OPEN
Microbial diversity drives multifunctionality in terrestrial ecosystems Manuel Delgado-Baquerizo1, Fernando T. Maestre2, Peter B. Reich1,3, Thomas C. Jeffries1, Juan J. Gaitan4, Daniel Encinar2, Miguel Berdugo2, Colin D. Campbell5 & Brajesh K. Singh1,6
Despite the importance of microbial communities for ecosystem services and human welfare, the relationship between microbial diversity and multiple ecosystem functions and services (that is, multifunctionality) at the global scale has yet to be evaluated. Here we use two independent, large-scale databases with contrasting geographic coverage (from 78 global drylands and from 179 locations across Scotland, respectively), and report that soil microbial diversity positively relates to multifunctionality in terrestrial ecosystems. The direct positive effects of microbial diversity were maintained even when accounting simultaneously for multiple multifunctionality drivers (climate, soil abiotic factors and spatial predictors). Our findings provide empirical evidence that any loss in microbial diversity will likely reduce multifunctionality, negatively impacting the provision of services such as climate regulation, soil fertility and food and fibre production by terrestrial ecosystems.
1 Hawkesbury Institute for the Environment, Western Sydney University, Penrith, New South Wales 2751, Australia. 2 A ´ rea de Biodiversidad y Conservacio´n, Departamento de Biologı´a y Geologı´a, Fı´sica y Quı´mica Inorga´nica, Escuela Superior de Ciencias Experimentales y Tecnologı´a, Universidad Rey Juan Carlos, Calle Tulipa´n Sin Nu´mero, Mo´stoles 28933, Spain. 3 Department of Forest Resources, University of Minnesota, St Paul, Minnesota 55108, USA. 4 Instituto de Suelos, CIRN, INTA, Nicolas Repetto y de los Reseros Sin Nu´mero, Hurlingham, Buenos Aires 1686, Argentina. 5 The James Hutton Institute, Craigiebuckler, Aberdeen AB15 8QH, UK. 6 Global Centre for Land-Based Innovation, Western Sydney University, Penrith South DC, New South Wales 2751, Australia. Correspondence and requests for materials should be addressed to M.D-B. (email: [email protected]).
NATURE COMMUNICATIONS | 7:10541 | DOI: 10.1038/ncomms10541 | www.nature.com/naturecommunications
1
ARTICLE
A
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10541
large body of research conducted during the past two decades indicates that ecosystem functioning is positively related to plant diversity1–4. Unlike plants, we have limited knowledge of the relationship between microbial diversity and ecosystem functioning, particularly in terrestrial environments5,6. Microbial communities play key roles in maintaining multiple ecosystem functions and services simultaneously (‘multifunctionality’ hereafter), including nutrient cycling, primary production, litter decomposition and climate regulation7–10. Experiments carried out under controlled conditions8,11–13 suggest that the diversity of soil organisms can promote multifunctionality. However, none of these studies have explicitly addressed the relationship between soil microbial diversity and multifunctionality on the global scale. Of the various ecosystem processes on Earth, plant productivity and nutrient cycling are among those most important for supporting human welfare9,14. Because of the continuous global population growth15, substantial increases in plant production and land use intensification will be required to support future demand for food and fibre14. Understanding the factors controlling the multiple functions linked to plant production and nutrient cycling under a changing environment is, thus, critical to preserve and manage natural and human-dominated ecosystems. We posit that soil microbial diversity plays a key role in maintaining ecosystem multifunctionality by supporting processes such as litter decomposition and organic matter mineralization3,7,16,17, which allow transfer of matter and energy between above- and belowground communities12,16–19. There is a growing body of experimental and observational studies providing evidence that the relationship between biodiversity (that is, microbes and plants) and ecosystem functioning is more linear than saturating6,20,21. Thus, any loss in microbial diversity as a consequence of global environmental changes such as land use, nitrogen enrichment and climate change3,9,10,14,22,23 would likely alter the capacity of microbes to sustain multiple above- and belowground ecosystem functions. However, we lack empirical evidence on the relationships between microbial diversity and multifunctionality in terrestrial ecosystems, and few studies have addressed the relative importance of this diversity versus other drivers of ecosystem functioning, such as soil abiotic properties, climate and plant species richness8,18. This hampers our ability to predict changes in multifunctionality under ongoing global environmental change, and to formulate sustainable management and conservation policies10. Here, we hypothesize that microbial diversity: (i) promotes multifunctionality in terrestrial ecosystems; and (ii) is as important as variables such as soil pH, climate and spatial predictors, latitude and altitude as drivers of variation in multifunctionality. We tested these hypotheses using data from two large-scale surveys, a global study including 78 drylands from all continents except Antarctica (‘Drylands’ hereafter)24,25 and a national soil survey including 179 locations in Scotland (‘Scotland’ hereafter)26. The Drylands data set include diverse ecosystem types (grasslands, mixed grassland/woodland and woodlands), and provides a wide range of environmental conditions typically found in drylands worldwide. Similarly, the Scotland data set includes six ecosystem types (bog, moorland, semi-natural grassland, forest, arable and improved grassland) covering the whole of Scotland, and is representative of many soil types and land uses found in northern temperate regions. Our intention is not to merge both data sets, which indeed have some differences in sampling design and experimental methods, but to test our hypotheses using two independent and large-scale data sets from ecosystems widely differing in their vegetation, climatic and soil attributes24–26. 2
We found that soil microbial diversity is positively related to multifunctionality in both the Drylands and the Scotland data sets. The positive effects of microbial diversity on multifunctionality were maintained even when accounting simultaneously for multiple climatic, abiotic and spatial predictors of multifunctionality. Our study provides empirical evidence that microbial diversity positively relates to multifunctionality in terrestrial ecosystems on the global scale; and further suggests that any loss in microbial diversity will likely reduce the rates at which multiple ecosystem functions and services are being maintained in terrestrial ecosystems. Results and Discussion Microbial diversity and ecosystem multifunctionality. A total of 166,244/24,249 (bacteria/fungi) and 49,102 (bacteria) operational taxonomic units (OTUs) were found in the Drylands and Scotland data sets, respectively (see Supplementary Fig. 1 for rarefaction curves and Supplementary Figs 2 and 3 for the dominant taxa found). We first explored the relationship between microbial diversity, estimated with the Shannon index (ref. 27), and multifunctionality, evaluated using the standardized average of six variables that were available for the two data sets: potential net nitrogen (N) mineralization, nitrate, ammonium, DNA concentration, available phosphorus (P) and plant productivity (see Methods). Soil microbial diversity positively relates to multifunctionality in both data sets (Fig. 1). These results were maintained when controlling for the spatial structure of the data by using spatial autoregressive analyses25,28 (Fig. 1). We also found positive relationships between soil microbial diversity and most of the individual functions measured, as well as between this diversity and most of the possible combinations among functions (Supplementary Table 1). Our multifunctionality index was also strongly related, for each data set, to an extended version of this index including 8 and 17 soil functions that were unique to the Scotland and Drylands data sets, respectively (Supplementary Figs 4 and 5). Further analyses provided evidence that Shannon diversity was positively and strongly related to biodiversity components such as phylogenetic diversity and species richness (Supplementary Figs 6–9). Phylogenetical diversity and species richness were also highly and positively related to multifunctionality (Supplementary Figs 6–9). Finally, our results were robust to the approach used to quantify multifunctionality: single functions (Supplementary Table 1), averaging multifunctionality (Fig. 1) and multiple-threshold multifunctionality (Supplementary Figs 10 and 11). The multiple-threshold approach provided additional evidence that the effect of microbial diversity in the number of functions surpassing different thresholds of functionality is mainly positive and significant (Supplementary Figs 10 and 11). Also, the maximum number of functions maximized is the same than the number of functions measured (six, see Supplementary Fig. 10), which indicates that there are no trades-offs between the functions evaluated in our study. Moreover, the multiplethreshold approach indicated that the effect of diversity over multifunctionality is moderate-high in the Drylands data set and high in the Scotland data set (see Supplementary Figs 10 and 11). In particular, these results indicate that microbial diversity in Drylands has a significant effect on the ability of the system to provide more functions working at moderate to high performance levels (peaks B40 and 60% in Supplementary Fig. 11a,b respectively), whereas in Scotland this is expanded to functions working at really high performance levels (peak B75% in Supplementary Fig. 11c). Albeit our results are correlative in nature, and hence cannot be taken as a definitive proof of causation, they agree with those from theoretical and
NATURE COMMUNICATIONS | 7:10541 | DOI: 10.1038/ncomms10541 | www.nature.com/naturecommunications
ARTICLE
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10541
a
Drylands 1.5 1.0
Grassland Mixed Woodland
c 2.0 1.5 1.0
0.5
Ecosystem multifunctionality (unitless)
0.0 –0.5
Scotland Bogs Moorland Woodland Semi-natural grassland Improved grassland Arable
0.5 0.0 –0.5 –1.0
–1.0
–1.5
–2.0 –1.5 10.0 10.5 9.5 11.0 11.5 12.0 8.0 9.0 5.0 6.0 7.0 10.0 Diversity of bacteria (shannon index, bits) Diversity of bacteria (shannon index, bits)
b
1.5 1.0 0.5 0.0
–0.5 –1.0 –1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 Diversity of fungi (shannon index, bits)
Figure 1 | Relationships between microbial diversity and ecosystem multifunctionality. Results are shown for the Drylands (bacteria (a) and fungi (b)) and Scotland (bacteria (c)) data sets. The solid and dashed lines represent the fitted ordinary least squares (OLS) and simultaneous autoregression (SAR) models, respectively. Results of regressions are as follows: (a) OLS, R2 ¼ 0.118, P ¼ 0.012, AICc ¼ 133.463; SAR, R2 ¼ 0.101, P ¼ 0.005, AICc ¼ 135.013; (b) OLS, R2 ¼ 0.235, P ¼ 0.002, AICc ¼ 122.399; SAR, R2 ¼ 0.215, Po0.001, AICc ¼ 124.433 (c) OLS, R2 ¼ 0.226, Po0.001, AICc ¼ 265.539; SAR, R2 ¼ 0.222, Po0.001, AICc ¼ 266.574.
experimental studies showing a positive relationships between overall soil diversity and multiple soil functions, such as those used here8,12,13,16,29,30. Moreover, a recent field observational study has also found positive relationships between bacterial diversity and multifunctionality in the Chinese Tibetan Plateau18. Our results provide, to our knowledge, the first empirical evidence showing that microbial diversity positively relates to multifunctionality in terrestrial ecosystems on the global scale. Consequently, our results support the hypothesis that microbial diversity can be critical to maintain multifunctionality8,18, suggesting that losses of microbial diversity will likely reduce the ability of terrestrial ecosystems to provide critical ecosystem services. Accounting for multiple multifunctionality drivers. We used Random Forest modelling31 to identify the most important predictors (distance from equator, altitude, mean annual temperature (MAT), mean annual precipitation (MAP), soil pH and microbial diversity) of multifunctionality; and structural equation modelling (SEM) (ref. 32) to test whether the relationship between microbial diversity and multifunctionality is maintained when accounting for multiple multifunctionality drivers simultaneously (see a priori model in Supplementary Fig. 12). Our Random Forest models indicate that microbial diversity was as important as or more important than other multifunctionality predictors (Fig. 2). Indeed, microbial diversity was more important than MAT and altitude in the two data sets, and than MAP in the Scotland data set (Fig. 2b). Similar results were found after including ecosystem type as a predictor in these analyses (Supplementary Fig. 13; see Supplementary Figs 14 and 15 for values of the functions measured across ecosystem types). The role of distance from equator, altitude, climate and soil pH as predictors of multifunctionality is well known17,25. Most relevant
to the topic of this study, we found that microbial diversity was a major predictor of multifunctionality in the two data sets used, even after accounting for the simultaneous direct and indirect effects of these variables (Fig. 2a,b). Our SEMs explained 53 and 38% of the variance found in the ecosystem multifunctionality of the Drylands and Scotland data sets, respectively (Fig. 3a,b). In both cases we found a direct positive effect of microbial diversity on multifunctionality (Fig. 3). In the Drylands data set, fungal diversity showed a slightly higher total positive effect than bacterial diversity on multifunctionality. Fungi are known to be more tolerant of desiccation than bacteria33, and thus fungal diversity may have a predominant effect on multifunctionality in drylands, where soils remain under dry conditions during most of the year34. Not surprisingly, the effects of climate and soil pH on multifunctionality followed opposite trends in the Drylands and Scotland data sets, as indicated by the standardized total effects from SEM (Fig. 3c,d). The biological activity and productivity of drylands are well known to be limited by rainfall, rather than by temperature (except in cold deserts)34; consistent with this, MAP and multifunctionality were positively related in our data set. Contrarily, temperature, but not MAP, is known to limit ecosystem functioning in mesic cold temperate ecosystems such as those from Scotland35. Similarly, soil pH is often basic in drylands (for example, because of carbonate accumulation) and acid in cold temperate ecosystems (for example, due to organic matter accumulation)35. Therefore, assuming microbes are adapted to the typical pH of their habitats, soil pH influences multifunctionality in a distinct manner in the two data sets studied. Despite these contrasting effects, the positive direct and total effects of microbial diversity on multifunctionality were always maintained, and were robust to the analytical methods used here (linear regression, random forest and SEM). Collectively, these results demonstrate that microbial diversity
NATURE COMMUNICATIONS | 7:10541 | DOI: 10.1038/ncomms10541 | www.nature.com/naturecommunications
3
ARTICLE
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10541
a
b
Drylands
Scotland
120 2
2
R = 0.42; P
DOI: 10.1038/ncomms10541
OPEN
Microbial diversity drives multifunctionality in terrestrial ecosystems Manuel Delgado-Baquerizo1, Fernando T. Maestre2, Peter B. Reich1,3, Thomas C. Jeffries1, Juan J. Gaitan4, Daniel Encinar2, Miguel Berdugo2, Colin D. Campbell5 & Brajesh K. Singh1,6
Despite the importance of microbial communities for ecosystem services and human welfare, the relationship between microbial diversity and multiple ecosystem functions and services (that is, multifunctionality) at the global scale has yet to be evaluated. Here we use two independent, large-scale databases with contrasting geographic coverage (from 78 global drylands and from 179 locations across Scotland, respectively), and report that soil microbial diversity positively relates to multifunctionality in terrestrial ecosystems. The direct positive effects of microbial diversity were maintained even when accounting simultaneously for multiple multifunctionality drivers (climate, soil abiotic factors and spatial predictors). Our findings provide empirical evidence that any loss in microbial diversity will likely reduce multifunctionality, negatively impacting the provision of services such as climate regulation, soil fertility and food and fibre production by terrestrial ecosystems.
1 Hawkesbury Institute for the Environment, Western Sydney University, Penrith, New South Wales 2751, Australia. 2 A ´ rea de Biodiversidad y Conservacio´n, Departamento de Biologı´a y Geologı´a, Fı´sica y Quı´mica Inorga´nica, Escuela Superior de Ciencias Experimentales y Tecnologı´a, Universidad Rey Juan Carlos, Calle Tulipa´n Sin Nu´mero, Mo´stoles 28933, Spain. 3 Department of Forest Resources, University of Minnesota, St Paul, Minnesota 55108, USA. 4 Instituto de Suelos, CIRN, INTA, Nicolas Repetto y de los Reseros Sin Nu´mero, Hurlingham, Buenos Aires 1686, Argentina. 5 The James Hutton Institute, Craigiebuckler, Aberdeen AB15 8QH, UK. 6 Global Centre for Land-Based Innovation, Western Sydney University, Penrith South DC, New South Wales 2751, Australia. Correspondence and requests for materials should be addressed to M.D-B. (email: [email protected]).
NATURE COMMUNICATIONS | 7:10541 | DOI: 10.1038/ncomms10541 | www.nature.com/naturecommunications
1
ARTICLE
A
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10541
large body of research conducted during the past two decades indicates that ecosystem functioning is positively related to plant diversity1–4. Unlike plants, we have limited knowledge of the relationship between microbial diversity and ecosystem functioning, particularly in terrestrial environments5,6. Microbial communities play key roles in maintaining multiple ecosystem functions and services simultaneously (‘multifunctionality’ hereafter), including nutrient cycling, primary production, litter decomposition and climate regulation7–10. Experiments carried out under controlled conditions8,11–13 suggest that the diversity of soil organisms can promote multifunctionality. However, none of these studies have explicitly addressed the relationship between soil microbial diversity and multifunctionality on the global scale. Of the various ecosystem processes on Earth, plant productivity and nutrient cycling are among those most important for supporting human welfare9,14. Because of the continuous global population growth15, substantial increases in plant production and land use intensification will be required to support future demand for food and fibre14. Understanding the factors controlling the multiple functions linked to plant production and nutrient cycling under a changing environment is, thus, critical to preserve and manage natural and human-dominated ecosystems. We posit that soil microbial diversity plays a key role in maintaining ecosystem multifunctionality by supporting processes such as litter decomposition and organic matter mineralization3,7,16,17, which allow transfer of matter and energy between above- and belowground communities12,16–19. There is a growing body of experimental and observational studies providing evidence that the relationship between biodiversity (that is, microbes and plants) and ecosystem functioning is more linear than saturating6,20,21. Thus, any loss in microbial diversity as a consequence of global environmental changes such as land use, nitrogen enrichment and climate change3,9,10,14,22,23 would likely alter the capacity of microbes to sustain multiple above- and belowground ecosystem functions. However, we lack empirical evidence on the relationships between microbial diversity and multifunctionality in terrestrial ecosystems, and few studies have addressed the relative importance of this diversity versus other drivers of ecosystem functioning, such as soil abiotic properties, climate and plant species richness8,18. This hampers our ability to predict changes in multifunctionality under ongoing global environmental change, and to formulate sustainable management and conservation policies10. Here, we hypothesize that microbial diversity: (i) promotes multifunctionality in terrestrial ecosystems; and (ii) is as important as variables such as soil pH, climate and spatial predictors, latitude and altitude as drivers of variation in multifunctionality. We tested these hypotheses using data from two large-scale surveys, a global study including 78 drylands from all continents except Antarctica (‘Drylands’ hereafter)24,25 and a national soil survey including 179 locations in Scotland (‘Scotland’ hereafter)26. The Drylands data set include diverse ecosystem types (grasslands, mixed grassland/woodland and woodlands), and provides a wide range of environmental conditions typically found in drylands worldwide. Similarly, the Scotland data set includes six ecosystem types (bog, moorland, semi-natural grassland, forest, arable and improved grassland) covering the whole of Scotland, and is representative of many soil types and land uses found in northern temperate regions. Our intention is not to merge both data sets, which indeed have some differences in sampling design and experimental methods, but to test our hypotheses using two independent and large-scale data sets from ecosystems widely differing in their vegetation, climatic and soil attributes24–26. 2
We found that soil microbial diversity is positively related to multifunctionality in both the Drylands and the Scotland data sets. The positive effects of microbial diversity on multifunctionality were maintained even when accounting simultaneously for multiple climatic, abiotic and spatial predictors of multifunctionality. Our study provides empirical evidence that microbial diversity positively relates to multifunctionality in terrestrial ecosystems on the global scale; and further suggests that any loss in microbial diversity will likely reduce the rates at which multiple ecosystem functions and services are being maintained in terrestrial ecosystems. Results and Discussion Microbial diversity and ecosystem multifunctionality. A total of 166,244/24,249 (bacteria/fungi) and 49,102 (bacteria) operational taxonomic units (OTUs) were found in the Drylands and Scotland data sets, respectively (see Supplementary Fig. 1 for rarefaction curves and Supplementary Figs 2 and 3 for the dominant taxa found). We first explored the relationship between microbial diversity, estimated with the Shannon index (ref. 27), and multifunctionality, evaluated using the standardized average of six variables that were available for the two data sets: potential net nitrogen (N) mineralization, nitrate, ammonium, DNA concentration, available phosphorus (P) and plant productivity (see Methods). Soil microbial diversity positively relates to multifunctionality in both data sets (Fig. 1). These results were maintained when controlling for the spatial structure of the data by using spatial autoregressive analyses25,28 (Fig. 1). We also found positive relationships between soil microbial diversity and most of the individual functions measured, as well as between this diversity and most of the possible combinations among functions (Supplementary Table 1). Our multifunctionality index was also strongly related, for each data set, to an extended version of this index including 8 and 17 soil functions that were unique to the Scotland and Drylands data sets, respectively (Supplementary Figs 4 and 5). Further analyses provided evidence that Shannon diversity was positively and strongly related to biodiversity components such as phylogenetic diversity and species richness (Supplementary Figs 6–9). Phylogenetical diversity and species richness were also highly and positively related to multifunctionality (Supplementary Figs 6–9). Finally, our results were robust to the approach used to quantify multifunctionality: single functions (Supplementary Table 1), averaging multifunctionality (Fig. 1) and multiple-threshold multifunctionality (Supplementary Figs 10 and 11). The multiple-threshold approach provided additional evidence that the effect of microbial diversity in the number of functions surpassing different thresholds of functionality is mainly positive and significant (Supplementary Figs 10 and 11). Also, the maximum number of functions maximized is the same than the number of functions measured (six, see Supplementary Fig. 10), which indicates that there are no trades-offs between the functions evaluated in our study. Moreover, the multiplethreshold approach indicated that the effect of diversity over multifunctionality is moderate-high in the Drylands data set and high in the Scotland data set (see Supplementary Figs 10 and 11). In particular, these results indicate that microbial diversity in Drylands has a significant effect on the ability of the system to provide more functions working at moderate to high performance levels (peaks B40 and 60% in Supplementary Fig. 11a,b respectively), whereas in Scotland this is expanded to functions working at really high performance levels (peak B75% in Supplementary Fig. 11c). Albeit our results are correlative in nature, and hence cannot be taken as a definitive proof of causation, they agree with those from theoretical and
NATURE COMMUNICATIONS | 7:10541 | DOI: 10.1038/ncomms10541 | www.nature.com/naturecommunications
ARTICLE
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10541
a
Drylands 1.5 1.0
Grassland Mixed Woodland
c 2.0 1.5 1.0
0.5
Ecosystem multifunctionality (unitless)
0.0 –0.5
Scotland Bogs Moorland Woodland Semi-natural grassland Improved grassland Arable
0.5 0.0 –0.5 –1.0
–1.0
–1.5
–2.0 –1.5 10.0 10.5 9.5 11.0 11.5 12.0 8.0 9.0 5.0 6.0 7.0 10.0 Diversity of bacteria (shannon index, bits) Diversity of bacteria (shannon index, bits)
b
1.5 1.0 0.5 0.0
–0.5 –1.0 –1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 Diversity of fungi (shannon index, bits)
Figure 1 | Relationships between microbial diversity and ecosystem multifunctionality. Results are shown for the Drylands (bacteria (a) and fungi (b)) and Scotland (bacteria (c)) data sets. The solid and dashed lines represent the fitted ordinary least squares (OLS) and simultaneous autoregression (SAR) models, respectively. Results of regressions are as follows: (a) OLS, R2 ¼ 0.118, P ¼ 0.012, AICc ¼ 133.463; SAR, R2 ¼ 0.101, P ¼ 0.005, AICc ¼ 135.013; (b) OLS, R2 ¼ 0.235, P ¼ 0.002, AICc ¼ 122.399; SAR, R2 ¼ 0.215, Po0.001, AICc ¼ 124.433 (c) OLS, R2 ¼ 0.226, Po0.001, AICc ¼ 265.539; SAR, R2 ¼ 0.222, Po0.001, AICc ¼ 266.574.
experimental studies showing a positive relationships between overall soil diversity and multiple soil functions, such as those used here8,12,13,16,29,30. Moreover, a recent field observational study has also found positive relationships between bacterial diversity and multifunctionality in the Chinese Tibetan Plateau18. Our results provide, to our knowledge, the first empirical evidence showing that microbial diversity positively relates to multifunctionality in terrestrial ecosystems on the global scale. Consequently, our results support the hypothesis that microbial diversity can be critical to maintain multifunctionality8,18, suggesting that losses of microbial diversity will likely reduce the ability of terrestrial ecosystems to provide critical ecosystem services. Accounting for multiple multifunctionality drivers. We used Random Forest modelling31 to identify the most important predictors (distance from equator, altitude, mean annual temperature (MAT), mean annual precipitation (MAP), soil pH and microbial diversity) of multifunctionality; and structural equation modelling (SEM) (ref. 32) to test whether the relationship between microbial diversity and multifunctionality is maintained when accounting for multiple multifunctionality drivers simultaneously (see a priori model in Supplementary Fig. 12). Our Random Forest models indicate that microbial diversity was as important as or more important than other multifunctionality predictors (Fig. 2). Indeed, microbial diversity was more important than MAT and altitude in the two data sets, and than MAP in the Scotland data set (Fig. 2b). Similar results were found after including ecosystem type as a predictor in these analyses (Supplementary Fig. 13; see Supplementary Figs 14 and 15 for values of the functions measured across ecosystem types). The role of distance from equator, altitude, climate and soil pH as predictors of multifunctionality is well known17,25. Most relevant
to the topic of this study, we found that microbial diversity was a major predictor of multifunctionality in the two data sets used, even after accounting for the simultaneous direct and indirect effects of these variables (Fig. 2a,b). Our SEMs explained 53 and 38% of the variance found in the ecosystem multifunctionality of the Drylands and Scotland data sets, respectively (Fig. 3a,b). In both cases we found a direct positive effect of microbial diversity on multifunctionality (Fig. 3). In the Drylands data set, fungal diversity showed a slightly higher total positive effect than bacterial diversity on multifunctionality. Fungi are known to be more tolerant of desiccation than bacteria33, and thus fungal diversity may have a predominant effect on multifunctionality in drylands, where soils remain under dry conditions during most of the year34. Not surprisingly, the effects of climate and soil pH on multifunctionality followed opposite trends in the Drylands and Scotland data sets, as indicated by the standardized total effects from SEM (Fig. 3c,d). The biological activity and productivity of drylands are well known to be limited by rainfall, rather than by temperature (except in cold deserts)34; consistent with this, MAP and multifunctionality were positively related in our data set. Contrarily, temperature, but not MAP, is known to limit ecosystem functioning in mesic cold temperate ecosystems such as those from Scotland35. Similarly, soil pH is often basic in drylands (for example, because of carbonate accumulation) and acid in cold temperate ecosystems (for example, due to organic matter accumulation)35. Therefore, assuming microbes are adapted to the typical pH of their habitats, soil pH influences multifunctionality in a distinct manner in the two data sets studied. Despite these contrasting effects, the positive direct and total effects of microbial diversity on multifunctionality were always maintained, and were robust to the analytical methods used here (linear regression, random forest and SEM). Collectively, these results demonstrate that microbial diversity
NATURE COMMUNICATIONS | 7:10541 | DOI: 10.1038/ncomms10541 | www.nature.com/naturecommunications
3
ARTICLE
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10541
a
b
Drylands
Scotland
120 2
2
R = 0.42; P
